Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
From Marine Ecology to Developmental Biology In Honour of Pierre Tardent (1927-1997) VIE MILIEU, 2002, 52 (4) : 189-199 STASIS, CHANGE, AND FUNCTIONAL CONSTRAINT IN THE EVOLUTION OF ANIMAL BODY PLANS, WHATEVER THEY MAY BE G.A. WRAY1, R.R. STRATHMANN2 1 Department of 2 Friday Harbor Biology, Box 90338, Duke University, Durham, NC 27708-033, USA Laboratories and Department of Zoology, University of Washington, 620 University Road, Friday Harbor, WA 98250, USA [email protected] [email protected] BODY PLAN BAUPLAN DEVELOPMENT FUNCTION CONSTRAINT EVOLUTION ABSTRACT. – The phrase “body plan” or “bauplan” has been used to mean (1) the characteristic features of a phylum or other taxon of high rank, (2) architectural features of animals (such as symmetry; modular units; types of body walls, body cavities, body openings, and body subdivisions; types of supporting structures; position and structure of organ systems), (3) traits characteristic of an especially invariant stage in a life history (phylotypic stage), or (4) patterns of gene expression that first indicate the development of regions of the body. Multiple meanings of bodyplan within one argument can be misleading, but under all four meanings, body plans of animals have changed after stasis for long periods and after stasis during divergence of other traits. Change in body plans is often associated with an identifiable change in a functional constraint. Examples include decreases in body size and changes in requirements for feeding or locomotion. These observations support the hypothesis that functional constraints contribute to stasis in body plans. There is evidence that ancestral developmental processes constrain directions of evolutionary changes in body plans. There is little evidence that developmental processes prevent changes in body plans, but evidence for developmental constraint is more difficult to obtain than evidence for functional constraint. PLAN DU CORPS “BAUPLAN” DÉVELOPPEMENT FONCTION CONTRAINTE ÉVOLUTION RÉSUMÉ. – L’expression “plan du corps” ou “plan de construction [bauplan]” a été employée dans le passé pour désigner (1) les caractéristiques d’un phylum ou d’un taxon de rang élevé, (2) les caractères architecturaux des animaux (exemples: la symétrie; les unités modulaires; les types de tégument, de cavités du corps, d’ouvertures du corps, et de subdivisions du corps; les types de structures de renforcement; la position et la structure des systèmes d’organes), (3) les propriétés d’un stade ontogénétique particulièrement stable (stade phylotypique), ou encore (4) des motifs d’expression de gènes qui indiquent déjà le développement des différentes régions du corps. Des significations multiples de « plan du corps » à l’intérieur d’une même argumentation peuvent générer des erreurs, mais dans toutes les conditions évoquées de (1) à (4), les plans du corps des animaux ont été modifiés après de longues périodes de stase et après des stases accompagnées d’une divergence au niveau d’autres traits. Une modification du plan du corps est souvent liée à un changement identifiable dans le contexte d’une contrainte fonctionnelle. Parmi les exemples, on peut citer la réduction de taille du corps et les modifications des conditions requises pour l’alimentation et la locomotion. Ces observations confortent l’hypothèse selon laquelle les contraintes fonctionnelles contribuent à la stase d’un plan du corps. Il y a de très bons indices montrant que les processus ancestraux du développement sont contraignants quant aux directions possibles des changements évolutifs de plans du corps. Il y a très peu d’indices pour que les processus du développement empêchent tout changement de plan du corps, mais les preuves d’une contrainte du développement sont plus difficiles à trouver que les indices de contraintes fonctionnelles. 190 WRAY GA, STRATHMANN RR INTRODUCTION The potential for confusion from multiple meanings of “body plan” The term body plan (often used as a synonym of ground plan and bauplan) has been used with quite different meanings. The use of “body plan” in studies of evolutionary stasis and change can produce mistaken inferences when disparate meanings are mixed within an argument. As an example, stasis in body plans of animals has been inferred from the fossil record because first appearances of taxa of high rank are early in the Phanerozoic fossil record rather than scattered through the Phanerozoic. Animal phyla with fossilization potential appear as fossils in the Cambrian and not subsequently. Similarly, many of the animal classes are known from the Cambrian or Ordovician. If all members of a phylum or class share the same body plan, then about 500 million years of stasis in body plans is implied. This apparent stasis has prompted the question: Why have no new body plans originated since the Cambrian ? The answer depends in part on the definitions of body plan that are used and how they are combined. Not all members of a phylum or class have the traits that are used to characterize that group. They are classified with it because of evidence that they belong to the same clade. If the traits of animals rather than the traits that are used to characterize clades as taxa are considered, then the answer is that new body plans have originated since the Cambrian. The straw-man premise for the question “Why have there been no new body plans ?” was flawed, but it must be admitted that many ancestral traits characterizing animal phyla and classes have been retained for more than 500 million years. Thus some traits do exhibit remarkable stasis. These observations prompt questions that are not so easily dismissed. Why do body plans remain unchanged over such long periods ? Given such extended stasis, what enables them to change ? Such questions are being answered by functional and developmental biologists, but to interpret the answers one must consider the differing uses of the term “body plan.” Diverse meanings of “body plan” 1. The body plan often means the distinguishing morphological characteristics of a phylum or class. Under this definition, it would not be surprising that the traits that comprise a body plan are ancient and conservative. In so far as the hierarchical arrangement of taxa represents a sequence of evolutionary divergences, the traits that characterize higher taxa must have originated early. To be useful as diagnostic characters for descendants within the clade, the traits must change rarely. For more than two centuries zoologists have been searching for these ancient and conservative traits and redefining the animal phyla, and they are not done yet (Nielsen et al. 1996). Higher level systematics of animals has been an enduring source of controversy, and many zoologists are hoping that molecular evidence will help resolve questions about homology and homoplasy of traits that have been used to characterize taxa of high rank. Thus, even under this definition of body plan, one can find aspects of body plans that have changed in descendant lineages. Some of the traits characterizing a phylum vary within the phylum. We recognize the animals as members of a phylum or class because some indication of relationship remains. The observation that there are no new phyla since the Cambrian does not imply that there are no new body plans. It only implies that the animals with good potential for fossilization that originated later can be assigned to one of the clades known from the Cambrian. Similar observations apply to animal classes, with few new ones since the Ordovician, and to other ancient animal clades given high taxonomic rank. Nevertheless, there does appear to be a phenomenon to be explained. Many traits have remained unchanged for the majority of animals in a phylum or class. Groups characterized by these traits, if possessing readily fossilized parts, can now be traced as separate clades back to the Ordovician or Cambrian. For this meaning of body plan, observations of stasis and change in body plans are observations on traits that are considered to be persistent synapomorphies (shared derived characters) of clades of ancient origin. These traits gained the attention of biologists because they are ancient and persistent. The bias in the biologists’ choice is for prolonged stasis. With choice from a large number of traits, what is the probability that some would be so persistent by chance alone ? An estimate of that probability might be based on a null hypothesis of random change in traits during speciation and extinction, but we shall not attempt such an estimate here. 2. Body plan can also refer to architectural features of animals, such as types of body cavities, body walls, types of epithelia (Rieger 1994), organ systems, and skeletal support. Also included are arrangement of parts, as in segmentation, symmetry, position of mouth relative to nerve cords, or colonial versus solitary habit. Sometimes cell lineages (as in spiralian development) or cell movements (as in schizocoely and enterocoely), or cell types (such as choanocytes) are included as features of body plans. When body plans are defined as features of body organization rather than traits characterizing higher taxa, biases in selection of traits are eliminated (Fitch & Sudhaus 2002). The traits do not depend on relationships among clades, and other changes qualify as changes in body plans. STASIS AND CHANGE IN ANIMAL BODY PLANS 3. The traits of a phylotypic stage (Sander 1983, Raff 1996) have been suggested as the definition of the body plan, at least for those phyla or classes for which a phylotypic stage can be identified. Attention to phylotypic stages arose from the observation that some intermediate stage of development has diverged less than both earlier and later stages. For vertebrates the pharyngula has been considered to be a phylotypic stage, despite its variation among taxa (Ballard 1976, Richardson 1995). Most generally, a life-history stage has been considered to be phylotypic if it is the least varying stage of development in the most inclusive clade. To our knowledge there have been no attempts at a quantitative application of these dual criteria. Should greater weight be given to inclusiveness or to constancy in the identification of phylotypic stages ? Phylotypic stages are a special case of persistent synapomorphies. Like other persistent synapomorphies, they are selected by biologists for their stasis out of a large and uncounted set of traits. One explanatory hypothesis suggested for such especially conserved stages is that necessary interactions among developing body parts constrain evolutionary changes in this stage of development (Raff 1996). 4. A new use of the phrase “body plan” is its application to patterns of gene expression in embryos. Of particular interest are genes that establish the organization of developing embryos, including those that pattern general architectural features of body plans (body axes, germ layers, etc.) and distinguishing characteristics of particular body plans (notochord, segmentation, etc.). Most of the genes that pattern these features in embryos encode regulatory proteins such as transcription factors and intercellular signaling systems (Gerhart & Kirschner 1997, Carroll et al. 2001). The phrase “body plan” is increasingly used in association with the embryonic expression patterns of these genes (Fig. 1). Although there is clearly a relationship between regulatory genes and body plan features, gene expression profiles are not the same set of traits that have been used to distinguish phyla or classes, and they do not include all the features that distinguish types of body organization. As discussed in a later section, the phylogenetic distribution of most embryonic gene expression patterns does not match specific phyla and classes particularly well. The term “body plan” seems to gain special significance for biologists when disparate meanings are combined in one argument. Confounding several meanings in one argument can produce a false impression of stasis. Not all traits that characterize higher taxa are architectural. Clearly not all architectural features of animals are useful in characterizing higher taxa. Homologous genes may be expressed in a similar pattern yet quite different structures result in subsequent development, and similar structures can develop via rather different 191 patterns of gene expression. By using the term body plan (or ground plan or bauplan) ambiguously, one can attribute properties of one set of traits to another. When the attributes of the traits that characterize taxa of high rank are applied to the architectural features of multicellular animals or to the patterns of expression of homologous genes, or vice versa, confusion is likely. Use of “body plan” with double or triple meanings can bias assessment of the frequency or causes of stasis and change. What one means by evolutionary stasis or change in “body plans” clearly depends on the traits considered to be part of a body plan. Nevertheless, for all of the meanings listed above, there are examples of changes in body plan following change in a functional constraint. In contrast, it is more difficult to demonstrate developmental constraint as a cause of stasis in body plans. In selecting examples, we examined correlates of change because they are easier to examine than correlates of stasis. We selected examples that met one of two criteria. For some cases, the inferred phylogeny, the distribution of traits, and the fossil record indicated that the change occurred after prolonged stasis. In other cases there was little evidence on the duration of stasis or time of the change, but differences in body plan within taxa of low rank indicated that a trait remained unchanged during the evolution of marked morphological disparity but nevertheless changed subsequently (the alternative and unparsimonious hypothesis for examples in this second category is that the trait represents the ancestral condition and traits in all other members of the clade were convergently derived from it). CHANGES IN FUNCTIONAL REQUIREMENTS One explanation for stasis is stabilizing selection, with the explanation of change being a change in functional requirements. The functional requirements that may account for stasis and change in body plans are varied. Here we emphasize ancient synapomorphies and architectural features of animals. Nutrition from symbionts A variety of free-living animals with chemoautotrophic symbiotic bacteria have lost the gut lumen and in some cases the mouth and anus. As a change in the construction of the body, the loss of a gut may be considered a change in body plan. Indeed the absence of a recognizable gut in the adult contributed to the Pogonophora being considered a phylum, although morphological and molecular ev- 192 WRAY GA, STRATHMANN RR idence now nest them within the annelids (George & Southward 1973, McHugh 2000). A gut has been lost within families or genera within a number of clades (Gustafson & Lutz 1992, Krueger et al. 1992). The distribution of gutlessness within clades indicates extended stasis in which a gut was present followed by loss of a gut. Such distributions of gutlessness are found among species of protobranch bivalves, oligochaetes, and nematodes (Giere & Langheld 1987, Fisher 1990, Ott et al. 1982). As an example, the bivalve Solemya reidi has a stomach rudiment as a larva but lacks a functional gut. At metamorphosis, the larva ingests its own test cells, but the material enters the lumen of the perivisceral cavity (Gustafson & Reid 1988). Nutrition by parasitism The evolutionary transition from a free living habit to parasitism can involve extensive changes in body plans as ancestral structures for feeding and locomotion are lost and other structures elaborated. An extreme example from the crustacean arthropods illustrates the magnitude of possible changes. Rhizocephalan barnacles are allied to other cirripede crustaceans (Høeg 1995), but as adults the rhizocephalans lack such arthropod traits as segmentation and jointed appendages and in some cases are colonial. The female develops from a vermiform slug of cells injected into the host (Glenner et al. 2000). It becomes an interna, a set of branching roots with an epithelium and a cellular core. The externa, the reproductive part of the female, develops from the system of roots and extends outside the host. In one group of rhizocephalans, multiple externas develop from a set of rootlets (Høeg & Lützen 1993), thus achieving a colonial construction, with tissue connections and nutrient exchange maintained among morphological individuals. Thus the arthropod body plan has been so modified that it has been possible for coloniality to evolve as a novel feature of the body plan. The male rhizocephalan develops into a gametogenic structure hyperparasitic on the female. The multicellular structures of postlarval rhizocephalans give no morphological evidence of relationship to arthropods. Nutrient content of eggs Loss of a feeding larval stage has occurred numerous times within diverse phyla (Thorson 1950, Strathmann 1978, Hanken 1992). The transition from feeding to non-feeding development is often accompanied by loss or reduction of larval structures and accelerated development of postlarval structures. Loss of the larval mouth is common and loss of the entire gut has occurred in bryozoans, with development of the gut delayed until metamorphosis. These are substantial changes in the larval body plan. In some cases the loss has occurred after a demonstrably long period of stasis (Wray 1996). Cases of facultatively feeding larvae demonstrate that nutrient rich eggs and independence of exogenous food have evolved prior to evolutionary changes in the larval body (Hart 1996). Such changes are in part the result of loss of a functional requirement for obtaining food during the larval stage and as such could result from release from stabilizing selection (Strathmann 1975), but some of the changes may result from selection for improved performance of the non-feeding larva in swimming (Emlet 1994) and more rapid development to competence for metamorphosis (Wray & Bely 1994) and therefore could result from new directional selection. The changes in size of eggs and larval morphology and accelerated development of postlarval structures are associated with changes in cell fates that are evident early in development (Wray & Raff 1990). Among the most dramatic developmental changes are the transitions from holoblastic cleavage to incomplete (meroblastic) early cleavages. Such changes have occurred in both directions with increases and decreases in size of eggs, as in arthropods. The processes establishing developmental fates of portions of the embryo can differ between syncytial embryos, in which nuclei share cytoplasm without diffusion barriers, and embryos composed of separate cells. Other nutritional change A radula is characteristic of most classes of molluscs and is a plesiomorphic trait for gastropods, but it has been lost independently in several lineages that diverged since the Paleozoic (Oliverio 1995). These animals have been assigned to families or genera in which other species have a radula and they are easily identified as gastropods, but they have lost part of the gastropod body plan. Losses may be concentrated in a clade in which the radula had become specialized for delivery of toxin and the losses may be associated with a shift in diet (Kantor & Sysoev 1989). Changes from microphagous suspension feeding to macrophagy have occurred in several clades, with the derived macrophagous animals often occupying habitats that are poorer in suspended food. Sponges in the Cladorhizidae have lost the aquiferous system and choanocytes that are used in suspensionfeeding and that are characteristic of sponges. These sponges trap larger prey, such as crustaceans. The condition is inferred to be derived rather than ancestral. “Such a unique body plan would deserve recognition as a distinct phylum, if these animals were not so evidently close relatives of Porifera. STASIS AND CHANGE IN ANIMAL BODY PLANS Their siliceous spicules resemble those in several families of poecilosclerid Demospongiae” (Vacelet & Boury-Esnault 1995). The small ctenophore Ctenella aurantia lacks the colloblasts characteristic of Ctenophora and also lacks the peripheral canals (Carré & Carré 1993). It appears to be a highly modified cydippid, with functional changes associated with exogenic cnidocysts and small size. Motility and habitat selection Notochord, dorsal nerve cord, and postanal tail are chordate features that have become restricted to a non-feeding larval stage in ascidians. Their functional role is thus restricted to habitat selection and perhaps other kinds of dispersal. This restricted role has preceded loss of the larval tail and thus most of the chordate body independently in several lineages. A restricted functional role preceded their loss. Taillessness appears to have evolved more times within molgulids than in other ascidians, which raises the possibility of a developmental predisposition for tail loss in molgulids and by implication a greater developmental constraint on tail loss in other ascidians (Huber et al. 2000). Nevertheless, mutations resulting in loss of most of the body axis are known for other chordates, as in the floating head and no tail mutants of zebra fish (Halpern 1997). Stasis in the vertebrate body axis has not been from an absence of mutations. Most of the body posterior to the head can be removed by a mutation, but whereas some ascidians thrive without axial or appendicular locomotion, no vertebrates have adopted a mode of life that permits such a loss (Strathmann 2000). Adult size Evolution of smaller adults is often associated with changes in body plans. The annelids offer examples of changes in several components of the annelid body plan. In marine annelids, smaller size is associated with an interstitial habit and can be associated with loss of parapodia, loss of setae, and acquisition of rings of cilia in the adult (Westheide 1985). Reduction in size can also be associated with an acoelomate condition (Smith et al. 1986, Fransen 1988). Such divergences in body plan occur within families, as in the Hesionidae and in the Dorvilleidae together with the allied Dinophilidae. Loss of a blood vascular system and changes in excretory systems are also associated with reduction in size. In the hesionid polychaetes, small species that are apparently derived from larger ones have lost the blood vascular system and reacquired protonephridium-like excretory organs (Westheide 1986). This trend conforms to a broad correlation among multicellular animals (Ruppert & Smith 1988). In animals with a blood-vascular system, 193 muscle-mediated filtration from blood vessels through podocytes into the coelomic cavity produces the filtrate. The filtrate is then modified as fluid passes through an open duct from the coelom to the outside. In contrast, in animals without a blood vascular system, cilia-mediated filtration of fluids occurs from the coelom into the excretory duct from coelom to the outside. The cilia extend into the duct, and fluid is filtered through a weir formed by extensions of cells at the inner ends of the ducts. Ruppert and Smith note that development of these protonephridia-like and metanephridia-like systems does not conform to germ lines, and they question homology among protonephridia. An example of consequences of divergent size of adults within a species is provided by Bonellia, an echiuran with large females and minute parasitic dwarf males (Schuchert 1990). The females have a blood vascular system and open ducts between coelom and the outside. The males lack a blood vascular system, and have protonephridia-like excretory organs. The excretory organs of the male do not develop from a larval excretory system but rather as a new structure. Body plans can diverge between sexes within a species in association with different functional constraints. Skeletons and requirements for support, defense, and density Accretionary growth of a shell, as opposed to molting, has evolved in several arthropods. A bivalved carapace (as in conchostracans) or sessile habit (as in barnacles) appear to have been the preconditions for this change. Presumably the zoologists who initially classified barnacles as molluscs considered a permanent shell and mantle cavity to be important parts of a body plan. A particular form of calcite skeleton is the most consistent distinguishing feature of living and fossil echinoderms. Each skeletal element is composed of a latticework of anastomosing rods, the whole element optically like a single crystal of calcite. Skeletal elements are absent in the Pelagothuriidae, a family of elasipod sea cucumbers, although they belong to a clade of undoubted echinoderms and their inferred sister group within the suborder has skeletal ossicles (Hansen 1975). This condition may be associated with their pelagic life at great depths. Reduced skeletons, of which this is an extreme, carry consequences for buoyancy, structural support, flexibility, and defense. Protection of embryos Protection of gastropod embryos is associated with especially long cell cycles for the cells that will form trochoblasts and apical ectoderm, and the specification of the mesentoblast and dorso-ventral body axis occurs when the embryo has fewer cells 194 WRAY GA, STRATHMANN RR (van den Biggelaar et al. 1997). As a consequence, trochoblasts and other ectodermal cells can depart from a radial arrangement of cell fates and diverge in differentiation when the embryo has fewer cells. For example, fewer cells form prototrochal cilia and some instead form the head vesicle in pulmonate gastropods (van den Biggelaar 1993). Protection has led to modified cell interactions and cell fates to modify the trochophore. In addition to changes in ancient synapomorphies and architectural features, these are changes in what could be considered to be a phylotypic stage and presumably also in gene expression in the cells that do or do not form trochoblasts in the divergent clades. Conclusion on functional constraints on ancient synapomorphies and body structures The foregoing examples suggest that stasis in body plans (as architectural features of animals) results from stabilizing selection for performance of certain activities. Change results from changes in selection on performance. The foregoing examples include change and stasis in body plans under other definitions as well, with different subsets concerning traits characterizing major groups, body architecture, and changes at nearly all stages, including some that could be candidates for phylotypic stages. Functional requirements are, by themselves, sufficient to account for many instances of stasis and change in traits that are considered parts of body plans, but support for a hypothesis of functional constraint does not in itself exclude developmental constraints in these or other instances. A hypothesis of functional constraint emphasizes that high performance of some function is maintained by stabilizing selection. Developmental processes may influence the effect of mutations on performance and thus affect selection for or against the change. Our examples simply indicate that changes in body plans are associated with changes in the need to perform certain identifiable tasks. We now turn to meanings of “body plan” that are explicitly developmental. CHANGES IN DEVELOPMENT The two definitions of body plans that are based on development produce a somewhat different perspective on stasis and why it might occur. Here too, however, changes in body plans are clearly possible in post-Cambrian time and are commonly associated with shifts in functional constraints. Phylotypic stages and body plans Only three clades within phyla have been proposed to possess phylotypic stages: vertebrates, insects, and sea urchins (Ballard 1976, Sander 1983, Raff 1996). Within these clades, the phylotypic stage is generally well conserved (although exceptions exist, as described in the next paragraph). Most phyla, however, seem to lack a phylotypic stage. There are few obvious phylum-wide similarities in the embryonic or larval development of (for instance) cnidarians, bryozoans, nemerteans, and platyhelminthes. Either these and most other phyla lack phylotypic stages, or those stages are evolutionarily labile. In either case, a body plan definition based on phylotypic stages does not match the perception that body plans have remained static since the Cambrian. Modifications in phylotypic stages Some modifications to the phylotypic stage have evolved within each of the three groups proposed to possess one. Waddington (1956) and Elinson (1987) pointed out that the phylotypic stage of vertebrates is not the beginning of development, but rather a point of convergence that is preceded by more disparate earlier stages and followed by more disparate later ones. Duboule (1994) and Raff (1996) called this the “developmental hourglass” to emphasize the many changes in early (i.e., prephylotypic) development, most of which are associated with changes in functional constraints on embryonic nutrition (Elinson 1987, Raff 1996). The famous phylotypic “pharyngula” of vertebrates is itself not as stereotypic as implied by Haeckel’s famous figures, with component features often appearing in different sequences (Richardson 1995). The phylotypic stages of insects (germ band) and sea urchins (adult rudiment and juvenile) (Emlet 2000) show similar differences in the order of events and relative size of component structures. Thus, simple modifications to the phylotypic stage are possible. In a few cases, the phylotypic stage has been more extensively modified. The best examples come from sea urchins and other echinoderms with essentially direct development, where formation of the adult body plan shows almost no similarities to that of other species with overall rather similar adult anatomy (e.g., McEdward 1992, Schatt & Feral 1996, Emlet 2000). Gene expression and body plans Anatomical and gene expression indices of body plans show rather different patterns of stasis and change. Indeed, if all we had to go on were com- STASIS AND CHANGE IN ANIMAL BODY PLANS parisons of gene expression, it seems unlikely that we would recognize the same clades of animals as most distinctive. Many regulatory genes and features of their expression are very widely shared among animal phyla (Slack et al. 1993, Gerhart & Kirschner 1997, Carroll et al. 2001). Enthusiastic assertions that all animal embryos are patterned in essentially the same manner with the same set of regulatory genes (DeRobertis & Sasai 1996, Holland 1999, Carroll et al. 2001) probably overstate the case (Davidson 2001, Wilkins 2002). Nonetheless, some striking similarities exist. Genes of the Hox complex pattern position along the anteroposterior axis in all phyla for which the relevant functional information exists (Gerhart & Kirschner 1997). This applies to phyla with body plans on anatomical grounds that are about as different as they come, including cnidarians, arthropods, echinoderms, and chordates (Davidson 2001). Another good example is homologous regulatory proteins that control the differentiation of muscle and neuronal cell types in the same broad set of phyla (Davidson 2001). Most published comparisons of regulatory gene expression among phyla have emphasized similarities rather than differences. Some features of embryonic gene expression are apparently restricted to particular body plans, however. One is the expression of functionally antagonistic domains of engrailed and wingless at future segment boundaries in a wide variety of arthropods (Patel et al. 1989, Brown et al. 1994b). Unfortunately, no data are yet available from other ecdysozoan phyla with metameric or segmented body organization. Superficial similarities in engrailed expression have been reported from other phyla (Wedeen & Wiseblat 1991, Holland et al. 1997), but these cases do not appear to involve segment patterning, since expression is restricted to only a few taxa and occurs too late to play a role in segmentation (Bely & Wray 2001, Davidson 2001). Another example of body plan-specific regulatory gene expression appears to be the expression of brachyury in the notochord of urochordates and chordates. Although this gene has rather different developmental roles in other phyla, its function in notochord differentiation is apomorphic for chordates and appears to be widespread within the phylum. Additional examples of regulatory gene expression that closely match anatomical body plan features are scarce. This may be due partly to insufficient information. The depth of phylogenetic sampling of embryonic gene expression does not approach that for anatomical data , and for most animal phyla, the expression of not even a single gene has been examined. Four phyla have been extensively studied (arthropods, nematodes, chordates, and echinoderms) and scattered information is available for several others (cnidarians, annelids, mollusks, platyhelminths, and hemichordates). With 195 data from more phyla and from more taxa within each phylum, a closer match between gene expression and anatomical body plans may emerge. For now, however, regulatory genes are primarily remarkable for how many similarities they reveal among phenotypically disparate phyla and for how many changes have evolved within phyla (see below). Changes in body-patterning gene expression Losses and modifications in the expression of body-patterning genes have occurred within phyla, and some are associated with life history transformations. A well-studied example involves the wasp Copidosoma, which is a parasitoid with polyembryonic development. Unlike other wasps and holometabolous insects in general, the earliest body-patterning genes are no longer expressed in this species, including the “gap” genes and those that establish the anteroposterior axis. Later in embryogenesis, however, the conserved network of segmentation and homeotic gene expression is instituted (Grbic & Strand 1998). Another clear example involves the urochordate genus Molgula, where reduced larval dispersal has evolved several times independently by loss of the tail (Huber et al. 2000). Genes that pattern the notochord, a characteristic chordate feature, are expressed in embryos of species with tailed larvae but not in those whose larvae lack tails (Swalla & Jefferey 1996). Although these losses of characteristic aspects of embryonic gene expression in Copidosoma and Molgula both evolved in conjunction with shifts in life history, this correlation is not always evident. Some other dramatic changes in the expression of body-patterning genes have evolved without any obvious connection to functional changes in development or body plan organization. Examples include the role of bicoid in anteroposterior patterning of embryos of Diptera but not other holometabolous insects (Stauber et al. 2000), the role of even-skipped in pair-rule patterning of holometabolous but not hemimetabolous insects (Patel 1994), and the loss of homeotic patterning functions of zen and fushi-tarazu in holometabolous insects (Brown et al. 1994a, Dawes et al. 1994, Falciani et al. 1996). DISCUSSION Evolutionary changes in body plans abound, under four different meanings of the term body plan. The changes support the hypothesis that stasis in body plans depends on functional requirements. When functional requirements change, widely conserved features with long periods of stasis can be 196 WRAY GA, STRATHMANN RR changed. Extensive modifications in development often accompany these changes. Some of our examples are post-Paleozoic changes in body plans. For others, the variation is found within extant taxa of sufficiently low rank that one can infer stasis for the body plan through repeated evolutionary divergences in many other features prior to the change in the body plan. Multiple parallel cases are known for changes in body plans. These replicated evolutionary events are associated with parallel changes in functional requirements. Examples include changes in nutrient content of eggs, changes in requirements for motility, changes in nutrition of adults associated with parasitism, and changes in size of adults. Thus functional constraints on body plans are well documented. We did not, however, tally instances in which there was a change in functional requirements for a body plan but little change in the body plan. The concentration of evolutionary loss of tails in molgulids is consistent with differing constraints on mutational “knock outs” in different ascidians. The comparative evidence on developmental processes is suggestive. Molgulids lack an ankyrinlike protein that is present in the cortex of eggs in other sampled ascidian families, and this trait might be a developmental precondition that results in loss of tails in molgulids more readily than in other families of ascidians (Huber et al. 2000). Some sea urchins with large eggs and no apparent requirement for larval feeding, like Brisaster, have retained the pluteus form whereas other echinoids with large eggs and no requirement for feeding have lost the pluteus form (Hart 1996, Wray 1996). Also, some sea urchins with a feeding pluteus larva require thyroxine from their diet for development of the echinus rudiment, but at least one sea urchin with non-feeding larval development produces thyroxine (Saito et al. 1998). Endogenous production of thyroxine could be a developmental precondition for the evolutionary loss of a pluteus larval form. If so, dependence on diet for thyroxine could be considered a developmental constraint on the evolutionary transition to a non-feeding larva. It therefore remains possible that developmental processes differ in clades in which there have been apparently similar changes in functional requirements but difference in whether there has been a change or no change in body plan. Although hypotheses of developmental constraint have not fared well thus far, differences among clades in incidence of stasis and change raise the possibility of differing developmental constraints in different clades. Such differences offer the possibility of testing hypotheses of constraint by the comparative method. Despite intense interest, body plan-constraining features of development have yet to be demonstrated. Instead, studies of the cis-regulatory elements involved in body patterning have indicated Fig. 1. – Number of articles published between 1975 and 2001 that were identified with the key words “body plan” and “blastomeres” from the data base Medline/PubMed. Open squares are total number each year that were obtained with “body plan” and filled circles are for those articles whose titles indicated that gene expression or pattern specification were part of the study. The latter account for most of the accelerating increase in use of “body plan.” The + symbols are numbers of articles obtained with “blastomeres,”which increased during the same period, but only linearly. remarkably great scope for evolutionary change (Carroll et al. 2001). This does not, however, rule out body plan-constraining features of development. It seems unlikely from current evidence that tissue interactions constrain body plan features. It remains possible that gene interactions do so. Technical advances that would allow this possibility to be tested have come into widespread use in model systems and are now being applied in a comparative context. An extreme hypothesis of developmental constraint on body plans would be that variation in the traits comprising a body plan is impossible. This would mean that there are no mutations that change the body plan. An alternative hypothesis is that stabilizing selection accounts for stasis in body plans. Although changes are possible, they result in poorer performance of some key activity and are therefore selected against. As a functional arrangement of body parts is improved, there may be a decreasing number of routes to further improvements that do not pass through a functionally inferior transitional state. In addition, as ecological “space” is filled by a variety of animals, the minimal functional requirements may become greater and possible transitions between body plans thereby restricted (Frazzetta 1970, Valentine 1973, Strathmann 1978). New features arise very rarely, not because preexisting ones are “optimal” solu- STASIS AND CHANGE IN ANIMAL BODY PLANS 197 REFERENCES Fig. 2. – A model, based on Sewall Wright’s metaphor of an adaptive landscape, that illustrates constraints on evolutionary change in body plans. A, When a new way of life is first entered, clumsy transitional stages can nevertheless be improvements over coexisting forms. Many adaptive peaks may occupied, or at least closely approached. B, In an occupied way of life, reaching new adaptive peaks is less likely because occupation of adaptive peaks has produced deeper maladaptive valleys and perhaps because the adaptive changes involved in reaching the peaks removed traits that provided the initial evolutionary flexibility. C, In a mass extinction, adaptive peaks temporarily disappear, but the valleys between peaks may nevertheless remain steep because of survivors’ traits. D, Even when adaptive peaks reappear, they may remain unoccupied. Adaptive body plans may not appear as before because survivors continue to produce deep valleys in the adaptive landscape and because some ancestral traits are no longer represented in the descendents (After Strathmann 1978). tions (pan-adaptationist hypothesis) but because new features are very unlikely to be better than preexisting ones. Moreover, local optima that were once attained may not be attained a second time. One such situation is modeled in Fig. 2). ACKNOWLEDGEMENTS. – This study was aided by the Friday Harbor Laboratories, NSF grant IBN 0113603 to RRS, NSF grant IBN 96346 to GAW, a meeting on evolution of body plans organized by R A Raff at the University of Indiana, discussions with numerous colleagues, and an anonymous reviewer. Ballard WW 1976. Problems of gastrulation: real and verbal. Bioscience 26: 36-39. Bely AE, Wray GA 2001. Evolution of regeneration and fission in annelids: insights from engrailed- and orthodenticle-class gene expression. Development 128: 2781-2791. Brown SJ, Hilgenfeld RB, Denell RE 1994a. The beetle Tribolium castaneum has a fushi tarazu homolog expressed in stripes during segmentation. Proc Natl Acad Sci (USA) 91: 12922-12926. Brown SJ, Patel NH, Denell RE 1994b. Embryonic expression of the single Tribolium engrailed homolog. Dev Genet 15: 7-18. Carré C, Carré D 1993. Ctenella aurantia, genre et espèce nouveaux de Cténophore tentaculé (Ctenellidae fam. nov.) méditerranéen sans colloblastes et avec ventouses labiales. Canadian J Zool 71: 1804-1810. Carroll SB, Grenier JK, Weatherbee SD 2001. From DNA to Diversity: Molecular Genetics and the Evolution of Animal Design. Blackwell Scientific, Malden, Massachusetts. Davidson EH 2001. Genomic Regulatory Systems: Evolution and Development. Academic Press, San Diego. Dawes R, Dawson I, Falciani F, Tear G, Akam M 1994. Dax, a locust Hox gene related to fushi-tarazu but showing no pair-rule expression. Development 120: 1561-1672. DeRobertis EM, Sasai Y 1996. A common plan for dorsoventral patterning in Bilateria. Nature 380: 37-40. Duboule D 1994. Temporal colinearity and the phylotypic progression: a basis for the stability of a vertebrate Bauplan and the evolution of morphologies through heterochrony. Development 1994 supplement: 135-142. Elinson R 1987. Changes in developmental patterns: embryos of amphibians with large eggs. In Development as an Evolutionary Process, RA Raff & EC Raff, editors. Liss, New York: 1-21 Emlet RB 1994. Body form and patterns of ciliation in nonfeeding larvae of echinoderms: functional solutions to swimming in the plankton? Amer Zool 34: 570-585. Emlet RB 2000. What is a juvenile sea urchin? A comparative and phylogenetic survey of post-metamorphic juveniles. Zygote 8: S40-S41. Falciani F, Hausdorf B, Schröder R, Akam M, Tautz D, Denell R, Brown S 1996. Class 3 Hox genes in insects and the origin of zen. Proc Natl Acad Sci (USA) 93: 8479-8484. Fisher CR 1990. Chemoautotrophic and methanotrophic symbioses in marine invertebrates. Rev Aquat Sci 2: 399-436. Fitch DHA, Sudhaus W 2002. One small step for worms, one giant leap for “Bauplan ?” Evol & Devel 4: 243246. Fransen M 1988. Coelomic and vascular systems. Microfauna Marina 4: 199-213. Frazzetta TH 1970. From hopeful monsters to bolyerine snakes. Amer Nat 104: 55-72. George JD, Southward EC 1973. A comparative study of the setae of Pogonophora and polychaetous Annelida. J Mar Biol Assoc UK 53: 403-424. 198 WRAY GA, STRATHMANN RR Gerhart J, Kirschner M 1997. Cells, embryos, and evolution. Blackwell Science, Malden, Massachusetts. 642 p. Giere O, Lanheld C 1987. Structural organization, transfer and biological fate of endosymbiotic bacteria in gutless oligochaetes. Mar Biol 93: 641-650. Glenner H, Høeg JT, O’Brien JJ, Sherman TD 2000. Invasive vermigon stage in the parasitic barnacles Loxothylacus texanus and L. panopaei (Sacculinidae): closing of the rhizocephalan life-cycle. Mar Biol 136: 249-257. Grbic M, Strand MR 1998. Shifts in the life history of parasitic wasps correlate with pronounced alterations in early development. Proc Natl Acad Sci (USA) 95: 1097-1101. Gustafson RG, Lutz RA 1992. Larval and early post-larval development of the protobranch bivalve Solemya velum (Mollusca: Bivalvia). J Mar Biol Assoc UK 72: 383-402. Gustafson RG, Reid RGB 1988. Larval and post-larval morphogenesis in the gutless protobranch bivalve Solemya reidi (Cryptodonta: Solemyidae). Mar Biol 97: 373-387. Halpern ME 1997. Axial mesoderm and patterning of the zebrafish embryo. Amer Zool 37: 311-322 Hanken J 1992. Life history and morphological evolution. J Evol Biol 5: 549-557. Hansen B 1975. Systematics and biology of the deep-sea holothurians. Galathea-Report 13: 1-262. Hart MW 1996. Evolutionary loss of larval feeding: development, form and function in a facultatively feeding larva, Brisaster latifrons. Evolution 50: 174187. Hodin J, Hoffman JR, Miner BG, Davidson BJ 2001. Thyroxine and the evolution of lecithotrophic development in echinoids. In Echinoderms 2000. Edited by M Barker. A A Balkema, Lisse: 447-452. Høeg JT 1995. The biology and life cycle of the Rhizocephala (Cirripedia). J Mar Biol Assoc UK 75: 517550. Høeg JT, Lützen J 1993. Comparative morphology and phylogeny of the family Thompsoniidae (Cirripedia, Rhizocephala, Akentrogonida), with descriptions of three new genera and seven new species. Zoologica Scripta 22: 363-386. Holland LZ, Kene M, Williams NA, Holland ND 1997. Sequence and embryonic expression of the amphioxus engrailed gene (AmphiEn): the metameric pattern of transcription resembles that of its segmentpolarity homolog in Drosophila. Development 124: 1723-1732. Holland ND, Holland LZ 1999. Amphioxus and the utility of molecular genetic data for hypothesizing body part homologies between distantly related animals. Amer Zool 39: 630-640. Huber JL, da Silva KB, Bates WR, Swalla BJ 2000. The evolution of anural larvae in molgulid ascidians. Seminars Cell Devel Biol 11: 419-426. Kantor YI, Sysoev AV 1989. The morphology of toxoglossan gastropods lacking a radula, with a description of new species and genus of Turridae. J Molluscan Studie 55: 537-549. Krueger DM, Gallager SM, Cavanaugh CM 1992. Suspension feeding on phytoplankton by Solemya velum, a symbiont containing clam. Mar Ecol Prog Ser 86: 145-151. McEdward LR 1992. Morphology and development of a unique type of pelagic larva in the starfish Pteraster tesselatus (Echinodermata: Asteroidea). Biol Bull 182: 177-187. McHugh D 2000. Molecular phylogeny of Annelida. Can J Zool 78: 1873-1884. Nielsen CN, Scharff N, Eibye-Jacobsen D 1996. Cladistic analyses of the animal kingdom. Biol J Linn Soc 57: 385-410. Oliverio M 1995. The systematics of the radula-less gastropod Clathromangelia (Caenogastropoda, Conoidea). Zoologica Scripta 24: 193-201. Ott JG, Rieger G, Rieger R, Enders F 1982. New mouthless interstitial worms from the sulfide system: symbiosis with prokaryotes. Mar Ecol 3: 313-333. Patel NH 1994. Developmental evolution – insights from studies of insect segmentation. Science 266: 581-590. Patel NH, Martin-Blanco E, Coleman KG, Poole SJ, Ellis MC, Kornberg TB, Goodman CS 1989. Expression of engrailed proteins in arthropods, annelids, and chordates. Cell 58: 955-968. Raff RA 1996. The Shape of Life. University of Chicago Press, Chicago. 520 p. Richardson MK 1995. Heterochrony and the phylotypic period. Devel Biol 172: 412-421. Richardson M, Smith PR 1988. The functional organization of filtration nephridia. Biol Rev 63: 231-258. Rieger RM 1994. Evolution of the “lower” Metazoa. In Early Life on Earth, Edited by S Bengston, Columbia Univ Press, New York: 475-488. Ruppert EE, Smith PR 1988. The functional organization of filtration nephridia. Biol Rev 63: 231-258. Saito M, Seki M, Amemiya S, Yamasu K, Suyemitsu T, Ishihara K 1998. Induction of metamorphosis in the sand dollar Peronella japonica by thyroid hormones. Develop Growth Differ 40: 307-312. Sander K 1983. The evolution of patterning mechanisms: gleanings from insect embryogenesis and spermatogenesis. In Development and Evolution. Edited by BC Goodwin, N Holder, CC Wylie. Cambridge University Press, Cambridge: 137-159. Schatt P, Féral J-P 1996. Completely direct development of Abatus cordatus, a brooding schizasterid (Echinodermata: Echinoidea) from Kerguelen, with description of perigastrulation, a hypothetical new mode of gastrulation. Biol Bull 190: 24-44. Schuchert P 1990. The nephridium of the Bonellia viridis male (Echiura). Acta Zoologica 71: 1-4. Slack JMW, Holland PWH, Graham CH 1993. The zootype and the phylotypic stage. Nature 361: 490-492. Smith PR, Lombardi J, Rieger RM 1986. Ultrastructure of the body cavity lining in a secondary acoelomate, Microphthalmus cf. listensis Westheide (Polychaeta: Hesionidae). J Morph 188: 257-271. Stauber M, Taubert H, Schmidt-Ott U 2000. Function of bicoid and hunchback homologous in the basal cyclorrhaphan fly Megaselia (Phoride). Proc Natl Acad Sci (USA) 97: 10844-10849. Strathmann RR 1975. Larval feeding in echinoderms. Amer Zool 15: 717-730. Strathmann RR 1978. The evolution and loss of feeding larval stages of marine invertebrates. Evolution 32: 894-906. STASIS AND CHANGE IN ANIMAL BODY PLANS Strathmann RR 1978. Progressive vacating of adaptive types during the Phanerozoic. Evolution 32: 907-914. Strathmann RR 2000. Functional design in the evolution of embryos and larvae. Seminars Cell Devel Biol 11: 395-402. Swalla BJ, Jeffery WR 1996. Requirement of the Manx gene for expression of chordate features in a tailless ascidian larva. Science 274: 1205-1208. Thorson G 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol Rev 25:1-45. Vacelet J, Boury-Esnault N 1995. Carnivorous sponges. Nature 373: 333-375. Valentine JW 1973. Evolutionary Palaeoecology of the Marine Biosphere. Prentice-Hall, Englewood Cliffs, New Jersey. 511 p. van den Biggelaar JAM 1993. Cleavage pattern in embryos of Haliotis tuberculata (Archaeogastropoda) and gastropod phylogeny. J Morphol 216: 121-139. van den Biggelaar JAM, Dictus WJAG, van Loon AE 1997. Cleavage patterns, cell-lineages and cell specification are clues to phyletic lineages in Spiralia. Cell Develop Biol 8: 367-378. Waddington CH 1956. Principles of Embryology. George Allen & Unwin, London, 510 p. Wagner GP, Misof BY 1993. How can a character be developmentally constrained despite variation in developmental pathways ? J Evol Biol 6: 449-455. 199 Westheide W 1985. The systematic position of the Dinophilidae and the archiannelid problem. In The Origins and Relationships of Lower Invertebrates. Edited by S Conway Morris, R Gibson, HM Platt. Oxford Univ Press, Clarendon. 310-326. Wedeen CJ, Weisblat DA 1991. Segmental expression of an engrailed-class gene during early development and neurogenesis in an annelid. Development 113: 805814. Westheide W 1986. The nephridia of the interstitial polychaete Hesionides arenaria and their phylogenetic significance (Polychaeta, Hesionidae). Zoomorphology 106: 35-43. Wilkins AS 2002. The Evolution of Developmental Pathways. Sinauer Press, Sunderland Massachusetts. Wray GA 1996. Parallel evolution of nonfeeding larvae in echinoids. Syst Biol 45: 308-322. Wray GA, Bely AE 1994. The evolution of echinoderm development is driven by several distinct factors. Development 1994 supplement: 97-106. Wray GA, Lowe CJ 2000. Developmental regulatory genes and echinoderm evolution. Syst Biol 49: 28-51. Wray WG, Raff RA 1990. Novel origins of lineage founder cells in the direct-developing sea urchin Heliocidaris erythrogramma. Devel Biol 141: 41-54. Reçu le 31 mai 2002; received May 31, 2002 Accepté le 19 septembre 2002; accepted September 19, 2002